Eliminating the expression of a gene provides researchers with information on the functions of the gene. This can be done at the protein level by inhibiting protein functions with specific inhibitors or at the RNA level by preventing the mRNA from being translated into protein. Traditional methods for RNA level inhibition include antisense oligonucleotide and ribozyme. These methods of suppressing gene expression also provide potential treatments for human diseases.
A recent method of silencing gene expression at the mRNA level, termed RNA interference or RNAi, has emerged to be a very powerful alternative to the previous technologies. It works by a largely unknown yet much more active mechanism than those of antisense. Double-stranded RNA (“dsRNA”) is involved in this post-transcriptional gene silencing, a naturally occurring phenomenon in plants and fungi (Cogoni and Macino, Curr Opin Microbiol, 6:657-62. 1999). When introduced into worms, flies, or early mouse embryos, dsRNA induces a cellular response that degrades the mRNA that shares the same sequence with one strand of the dsRNA (Fire, Trends Genet, 9:358-363, 1999). In some systems, a few copies of the dsRNA can induce total degradation of target mRNAs (Fire et al., Nature, 6669:806-811, 1998). RNAi works at a very high success rate with almost any sequence in mRNAs (Caplen et al., Proc. Natl. Acad. Sci. USA 17:9742-9747, 2001).
RNAi holds great promise as a gene function study tool, and a potential therapeutic treatment for human diseases if it can be successfully used in mammalian cells or organisms. RNAi in most mammalian systems had been largely unsuccessful until recently because introduction of dsRNA similar to those used in worms and flies induced a general blockage of gene expression by mammalian anti-viral system involving PKR and interferon (Caplen et al., Gene, 1-2:95-105. 2000; Oates et al., Dev Biol, 1:20-8. 2000). However, by using a short, 21-23 nucleotide dsRNA, Elbashir et al. and other researchers showed that small interfering RNA (siRNA) could reduce or knock down specific gene expression without causing a global shut-down (Caplen et al., Proc. Natl. Acad. Sci. USA 17:9742-7. 2001; Elbashir et al., Nature, 6836:494-8. 2001).
The use of pol III promoters can effectively result in the production of siRNA inside mammalian cells. For example, an H 1-RNA promoter, which, like the U6 promoter, is a class III promoter of pol III, was used on a plasmid vector to direct transcription of a short hairpin RNA that was subsequently processed inside cells to produce siRNAs (Brummelkamp et al., Science 296:550-553. 2002). Similarly, the U6 promoter was used to produce short hairpin RNAs (Paddison et al., Genes Devel. 8:948-958. 2002; Paul et al., Nat. Biotechnol. 5:505-508. 2002; Sui et al., Proc. Natl. Acad. Sci. USA 8:5515-20. 2002; Yu et al., Proc. Natl. Acad. Sci. USA 9:6047-6052. 2002), or sense and antisense siRNAs (Lee et al., Nat. Biotechnol. 5:500-505, 2002; Miyagishi and Taira, Nat. Biotechnol. 5:497-500, 2002; Yu et al., Proc. Natl. Acad. Sci. USA 9:6047-52, 2002). Short hairpin RNAs (shRNAs) can mimic the naturally occurring micro-RNAs (miRNAs), which may be related to RNAi. However, current data concerning the design and processing of such artificial shRNAs to function as RNAi inducer remain conflicting. For example, the size (e.g. longer than 7 nucleotide) and sequence of the hairpin loop was found to be very important in one report (Brummelkamp et al., supra,. 2002), while shorter loops (1, 4 or 6 nucleotides) were used successfully by others (Paddison et al., supra, 2002; Paul et al., supra, 2002; Sui et al., supra, 2002; Yu et al., supra, 2002). A gene silencing effect was dependent on the order or orientation of the sense and the antisense strands within the hairpin in some cases (Paddison et al., supra, 2002), but marginally important or irrelevant in others (Paul et al., supra, 2002) (Yu et al., Proc. Natl. Acad. Sci. USA 9:6047-52. 2002). All of the above studies took advantage of the short class III of pol III promoters and its simple terminator to generate high copy number of short transcripts. A simple natural terminator might be insufficient to stop all transcripts at the desired position, though, as longer transcripts that have the potential of inducing non-specific expression shut down were observed (Lee et al., supra, 2002). Synthetic siRNAs or shRNAs transcribed from U6 promoters were effective in silencing transgenes or viral genes after being delivered by a hydrodynamic transfection method into adult mice (McCaffrey et al., Nature, 418:38-39, 2002).
The method of introducing into cells either long or short dsRNA molecules, isolated from in vitro or in vivo transcription, processing in soluble cell extracts, or chemical synthesis, has serious limitations. One limitation is that isolated dsRNA-mediated gene silencing, also referred to as gene knock-down, is only temporary in the sense that the effects of dsRNA can only last a limited number of cell divisions because the genetic information carried on RNA molecules is not integrated into the chromosome of the host cell. Another limitation is that RNA molecules are notoriously unstable and subject to degradation by environmentally abundant RNases. Therefore, handling RNA molecules demands extreme caution. Use of RNA molecules for therapeutic purpose has been unpractical. Modified RNA, such as adding protection groups to the nucleotides or changing the backbone of the polynucleotide chain, can give improved stability as in many cases involving antisense studies. However, several forms of modified RNA molecules that are more resistant to RNase degradation than natural RNA appeared to have lessened or lost RNAi capability, whereas replacing one strand of dsRNA with DNA did not induce RNAi at all (Parrish et al., Mol. Cell. 5:1077-87. 2000). In addition, the expenses for making RNA molecules are high and the process tedious and complicated.
As a result of these limitations, there is great interest in providing materials and methods for gene silencing that can be permanently introduced into cells or organisms (e.g. mammals such as a mouse or a human), and/or does not rely on using RNA molecules as mediators. DNA molecules, on the other hand, are more stable and cost-effective than RNA molecules. DNA molecules can be integrated into a host cell genome and can have long-term effect on the host cell. DNA molecules are relatively easy to synthesize and manipulate.